Saudi $867m Battery Storage Health & Degradation Monitor
Operating $867m worth of battery energy storage facilities requires predicting Lithium-ion cell State-of-Health (SoH) degradation over thousands of charge-discharge cycles. Build an ML degradation monitoring model using Scikit-Learn and Python to forecast capacity loss based on operational temperature and C-rate stress. Maintain a Mean Absolute Error (MAE) under 1.5% SoH.
What you are building
Develop a battery health (SoH) prediction model using Scikit-Learn and Python for Saudi BESS facilities.
How work is evaluated
Evaluates battery State of Health (SoH) prediction accuracy against test cell degradation profiles.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
MAE Error Threshold Check
Ensures absolute prediction error is less than 1.5% SoH.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Mean Absolute Error (SoH)
Percentage error in State-of-Health prediction. • target: 0.8 • range: 0-10
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Preprocess battery cell sensor telemetry (voltage, current, temperature, charge cycle count)
Train Random Forest / Gradient Boosting regression models on battery degradation curves
Predict remaining useful life (RUL) under desert thermal conditions
Output operational thermal cooling recommendations to slow degradation
Resources
Lithium-ion battery cell cycle degradation datasets recorded under high ambient temperature conditions.
Challenge setup
Evaluates battery State of Health (SoH) prediction accuracy against test cell degradation profiles.
Challenge input
JSON object with cycle_count, avg_temp_c, max_crate
Scikit-Learn
Standard Python ML toolkit for tabular predictive modeling.
Evaluated output
JSON object with predicted_soh_pct, mae_pct
- Ensures absolute prediction error is less than 1.5% SoH.
- Percentage error in State-of-Health prediction. • target: 0.8 • range: 0-10
- Benchmark: BatteryBench
- Mean Absolute Error (SoH) target: 0.8
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- Aim · Required
- MLflow · Optional
- Fixie · Optional
Evaluation contract
- MAE Error Threshold Check · Weight 1
- Mean Absolute Error (SoH) · Weight 1
Metrics
- Mean Absolute Error (SoH) · Target 0.8
Benchmarks
- BatteryBench
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with cycle_count, avg_temp_c, max_crate
Output: JSON object with predicted_soh_pct, mae_pct
Execution harness: Python
Evaluation tools
- Scikit-Learn · Machine Learning Library
Recipe freeze marker
No recipe freeze time is recorded.